Re: [cp-patches] Patch PING: java.util.{Currency, Date} and collection classes updates

[Michael Koch]

Am Donnerstag, 21. Oktober 2004 10:26 schrieb Andrew John Hughes:
> On Wed, 2004-10-20 at 07:39, Michael Koch wrote:
> > Am Mittwoch, 20. Oktober 2004 00:49 schrieb Andrew John Hughes:
> > > Hi all,
> > >  I still have a few patches waiting to be committed from a fair
> > > while back.  These are:
> > >
> > > * Revision 01: Documentation for java.util Collections
> > > http://lists.gnu.org/archive/html/classpath-patches/2004-09/msg
> > >0000 4.html
> >
> > +     * @return "[]".
> >
> > Should perhaps be '@return the string "[]"'
> >
> > true and false should be written as "<code>true</code>" to make
> > it look a little different in final documentation.
> >
> > Otherwise very good. Please commit.
>
> Thanks.  Committed.
>
> > >* Documentation and serialization patch for java.util.Currency
> > > http://lists.gnu.org/archive/html/classpath-patches/2004-09/msg
> > >0000 3.html
> >
> > +    for (int i=0;i<allLocales.length;i++)
> >
> > Please add whitespace before and after "=" and "<" and before
> > "i++". This makes it much more readable.
> >
> > Otherwise very good. Please commit.
>
> Must have missed that one -- I agree it looks ugly and I'm not the
> culprit.  Committed.
>
> > >* Documentation Patch 02 (java.util.Date)
> > > http://lists.gnu.org/archive/html/classpath-patches/2004-09/msg
> > >0002 4.html
> >
> > Okay, but please add a @specnote about when this was changed.
>
> Can you be a bit more specific here?  Are we talking about a
> specific change or something general to the class?  Its a while
> since I wrote this one, and all I can see is doc. updates.

I meant when was the "transient" keyword added ? or was it always part 
of the API ? If it was added later it would be good to add something 
like this into the javadoc:

@specnote It was made transient in JDK 1.3

(this is only an example)

> > > There is also:
> > >
> > > http://lists.gnu.org/archive/html/classpath-patches/2004-09/msg
> > >0005 9.html
> > >
> > > which really needs a Mauve test case before committal.
> >
> > The documentation parts are okay. Please commit them and provide
> > an extra patch only containing the functional changes for easier
> > review (perhaps together with mauve tests ;-)
>
> Will do -- it was dumb to bundle both in the first place.  What
> started out as a mere docs check turned into functional alterations
> as well. BTW, does anyone have a good suggestion of how to use
> Mauve when testing serialisation?   I know the JAPI stuff has a
> tool for testing this, but what is the preferred way from within
> Mauve?

Sorry, I never really touched serialization


Michael
-- 
Homepage: http://www.worldforge.org/